[xsdu]

There are faster implementations of tortoise that allow fine-tuning. You can get close to ElevenLabs quality if you have a perfect dataset. https://git.ecker.tech/mrq/ai-voice-cloning